diff --git a/common/lib/elf.c b/common/lib/elf.c
index 73c19979..702c95a1 100644
--- a/common/lib/elf.c
+++ b/common/lib/elf.c
@@ -577,7 +577,7 @@ again:
         }
 
 #if defined (__aarch64__)
-        clean_inval_dcache_poc(mem_base, mem_base + mem_size);
+        clean_dcache_poc(mem_base, mem_base + mem_size);
         inval_icache_pou(mem_base, mem_base + mem_size);
 #endif
     }
diff --git a/common/sys/cpu.c b/common/sys/cpu.c
index 7aabbbdc..90182abe 100644
--- a/common/sys/cpu.c
+++ b/common/sys/cpu.c
@@ -24,5 +24,6 @@ extern void delay(uint64_t cycles);
 extern size_t icache_line_size(void);
 extern size_t dcache_line_size(void);
 extern void clean_inval_dcache_poc(uintptr_t start, uintptr_t end);
+extern void clean_dcache_poc(uintptr_t start, uintptr_t end);
 extern void inval_icache_pou(uintptr_t start, uintptr_t end);
 extern int current_el(void);
diff --git a/common/sys/cpu.h b/common/sys/cpu.h
index 61ce5e5d..42352f56 100644
--- a/common/sys/cpu.h
+++ b/common/sys/cpu.h
@@ -265,6 +265,19 @@ inline void clean_inval_dcache_poc(uintptr_t start, uintptr_t end) {
     asm volatile ("dsb sy\n\tisb");
 }
 
+// Clean D-Cache to Point of Coherency
+inline void clean_dcache_poc(uintptr_t start, uintptr_t end) {
+    size_t dsz = dcache_line_size();
+
+    uintptr_t addr = start & ~(dsz - 1);
+    while (addr < end) {
+        asm volatile ("dc cvac, %0" :: "r"(addr) : "memory");
+        addr += dsz;
+    }
+
+    asm volatile ("dsb sy\n\tisb");
+}
+
 // Invalidate I-Cache to Point of Unification
 inline void inval_icache_pou(uintptr_t start, uintptr_t end) {
     size_t isz = icache_line_size();